GNGTS 2015 - Atti del 34° Convegno Nazionale

GNGTS 2015 S essione 1.1 35 in strike-slip and contractional settings and have a higher b-value of the Gutenberg-Richter power law. The elastic rebound is commonly considered as the main model for earthquake generation, being inferred as the mechanism dissipating the elastic energy accumulated during the interseismic period. This is likely true for contractional and strike-slip tectonic settings, but in tensional environments, the influence of gravitation may rather be dominant. In this paper we discuss some basic parameters that control the energy dissipation in shallow crustal extensional settings, such as the involved volume, the dip of the normal fault and the static friction. Natural examples will be taken from theApennines belt and other geodynamic settings, characterized by widespread extensional fault activity and related earthquakes. In addition we address the energy partitioning of earthquakes comparing the potential energy stored by the volume involved during the coseismic collapse with that inferred from earthquake magnitude. Regardless its origin (elastic or gravitational), potential energy has been demonstrated far greater, indicating that, in the energy budget, the available energy is far larger than that released by earthquakes waves. Therefore most of the energy must be dissipated by other geological phenomena (shear heating, heat flow and fracturing above all), consistently with previous works. It was also shown that all earthquakes gradually decrease the global gravitational energy, which is transformed into heat flow. However, as intuitively expected, in extensional tectonic settings the gravitational potential energy is decreasing, whereas it is increasing in contractional tectonic environments. In an upper crust having average density of 2.5 g/cm 3 , the vertical (lithostatic) load increases by about 25 MPa/km. Below 1–1.5 km depth, the crust is under horizontal compression even in extensional settings, i.e., the sigma 1 is vertical and corresponds to the lithostatic load, whereas the sigma 2 and 3 stress axes are horizontal but still contracting rocks. Therefore, say at 10 km depth, both sigma 2 and 3 must be positive and lower than 250 MPa. With the progression of the stretching, sigma 2 and 3 decrease, providing a larger differential stress that may eventually evolve into rupture and fault activation. This determines the collapse of the normal fault hangingwall, dissipating mostly gravitational potential energy, being the elastic component a minor factor in the fall, if any. In a simplified two-layers stretching crust, within the brittle upper crust, the faults are generally locked or slowly creeping, and the deformation is mostly assumed to be stick-slip. During the secular interseismic period of lithospheric stretching, the ductile lower crust is permanently shearing and thinning by viscous flow and deformation is inferred as a continuous process. The brittle-ductile transition (BDT) is on average located in the middle Earth’s crust. When a brittle fault merges into a ductile shear zone crosscutting the whole crust, the BDT is characterized by a pressure gradient because the brittle upper crust is mostly locked, whereas the viscous-plastic lower crust is sheared steadily (Fig. 1). Since the steady deformation of the ductile lower crust has to be transferred upward, it was proposed that the stretching could be accommodated in the brittle realm by dilation in a wedge conjugate to the episodically active normal fault (Fig. 1). In that wedge, millimetric open fractures are inferred to develop. These fractures may be partly filled by cement, and partly by fluids, as shown in logs of hydrocarbon exploration boreholes and as predicted by analogue modelling. This mechanism has also been defined as dilatancy, i.e., the phenomenon in which fractures and cracks form and open when rocks are stressed. This dilated wedge is inferred due to the strain partitioning and the pressure gradient between the ductile lower crust and the brittle upper crust. The occurrence of the dilated wedge from the geological model is also supported by the fact that the hangingwall of a normal fault could not collapse without a corresponding vacuity at the base of the activated fault segment (Fig. 1). A 10–15 km thick brittle crust needs only about 100–150 MPa to fail under extension. Moreover, once rocks are broken (e.g., by fracturing in limestone), they lose their elastic component, and fractures may be filled or unfilled by cement (Fig. 1), depending on fluids circulation, carbonate compensation depth, temperature, pressure, CO2 content in the system, etc. During the initial stage of collapse, fluid pressure increases, supporting the existence, in pre-coseismic stages, of open fractures filled by fluids, becoming squeezed by the fall of the fault hangingwall.